Origin of water in the terrestrial planets

Page created by Michele Gomez
 
CONTINUE READING
Origin of water in the terrestrial planets
Meteoritics & Planetary Science 40, Nr 4, 1–crossref to last page (2005)
           Abstract available online at http://meteoritics.org

                                   Origin of water in the terrestrial planets

                                                         Michael J. DRAKE

 Lunar and Planetary Laboratory, University of Arizona, Space Sciences Building, P. O. Box 210092, Tucson, Arizona 85721–0092, USA
                                                    E-mail: drake@lpl.arizona.edu

                                     (Received 19 January 2005; revision accepted 11 March 2005)

             Abstract–I examine the origin of water in the terrestrial planets. Late-stage delivery of water from
             asteroidal and cometary sources appears to be ruled out by isotopic and molecular ratio
             considerations, unless either comets and asteroids currently sampled spectroscopically and by
             meteorites are unlike those falling to Earth 4.5 Ga ago, or our measurements are not representative of
             those bodies. However, the terrestrial planets were bathed in a gas of H, He, and O. The dominant gas
             phase species were H2, He, H2O, and CO. Thus, grains in the accretion disk must have been exposed
             to and adsorbed H2 and water. Here I conduct a preliminary analysis of the efficacy of nebular gas
             adsorption as a mechanism by which the terrestrial planets accreted “wet.” A simple model suggests
             that grains accreted to Earth could have adsorbed 1–3 Earth oceans of water. The fraction of this water
             retained during accretion is unknown, but these results suggest that examining the role of adsorption
             of water vapor onto grains in the accretion disk bears further study.

                     INTRODUCTION                                       evidence for aqueous alteration in some carbonaceous
                                                                        chondrites.
     We take water for granted. Earth demonstrably has water                 There is, however, no consensus on the origin of water in
(Fig. 1). Using the Gamma Ray and Neutron Spectrometer                  the terrestrial planets. Suggested sources include comets,
(GRS) instrument aboard the Mars Odyssey spacecraft,                    hydrous asteroids, phyllosilicates migrating from the asteroid
Boynton et al. (2002) have found elevated H abundances that             belt, and hydrous minerals forming in the inner solar system
are most plausibly interpreted as vast water ice sheets at least        and accreting directly to the terrestrial planets. I examine each
1 m thick and buried under a thin layer of dust polewards of            of these hypotheses below. There are potential problems with
about 60° of latitude in both hemispheres of Mars (Fig. 2).             all of these sources. I propose that adsorption of water
And, of course, the residual polar caps are water ice. Carr             molecules from the gas in the accretion disk directly onto
(1996) presents evidence for widespread water erosion on                grains prior to their accretion into planetesimals and,
Mars. Further, Baker et al. (1991) present photogeological              eventually, planets may have been the source of some or all of
evidence for a boreal ocean on Mars. The D/H ratio of Venus             the water present in Venus, Earth, and Mars 4.5 Ga ago.
is about 100 times that of Earth’s oceans (Donahue and
Pollack 1983). We do not know Venus’ original water                                       SOURCES OF WATER
abundance and isotopic composition because Venus appears
to have lost most of its water in a way that strongly                        The origin of water in the terrestrial planets is intimately
fractionated the D/H ratio. One plausible explanation is                associated with the nature of their “building blocks.” There
preferential loss of H relative to D through UV                         are two end-member possibilities. One possibility is that
photodissociation of H2O at the top of the Venusian                     temperatures were too high in the inner solar system for
atmosphere caused by solar UV light. Mercury and the Moon               hydrous phases to exist in the accretion disk, so the terrestrial
appear to be bone-dry, possibly due to volatile loss in giant           planets accreted “dry” (Boss 1998). Water and probably
impacts (Taylor 2001), though both may have water ice                   organics were delivered from exogenous sources (e.g.,
trapped in permanently shadowed craters around their poles              comets, meteorites) after the terrestrial planets had formed. In
(Harmon 1997). Asteroids appear to be both hydrous and                  this case, all terrestrial planets should have had the same
anhydrous bodies, with the hydrous bodies being                         source of water, unless the delivery process was stochastic
concentrated towards the outer part of the Main Belt (Gradie            and dominated by one or a few large impacts of hydrous
et al. 1989), and McSween (1979) presents compelling                    objects.

                                                                    1                        © The Meteoritical Society, 2005. Printed in USA.
Origin of water in the terrestrial planets
2                                                            M. J. Drake

Fig. 1. The Atlantic Ocean on Earth. A view of the Copacabana Beach from the Sugar Loaf Mountain, Rio de Janeiro, September 2004. The
Meteoritical Society’s meeting was held at a hotel at the far end of the beach.

     At the other extreme is the possibility that the terrestrial    Comets
planets accreted “wet,” with anhydrous and hydrous silicate
phases among the material accreted to the growing planets                 Comets were long considered the leading candidate for
(Drake and Righter 2002). In this view, planetary water has          the origin of water in the terrestrial planets. This hypothesis
an indigenous origin. In this case, all terrestrial planets          was attractive for two reasons. First, it is widely believed that
probably also had the same source of water, unless hydrous           the inner solar system was too hot for hydrous phases to be
phases were stable at the orbit of, say, Mars but not closer to      thermodynamically stable (Boss 1998). Thus an exogenous
the Sun.                                                             source of water was needed. Second, the Earth and other
     Distinguishing between these end-members, and even              terrestrial planets underwent one or more magma ocean
variants of one, requires an understanding of the abundance of       events (Righter and Drake 1996; Li and Agee 1996; Righter
water and the oxygen and hydrogen isotopic composition of            and Drake 1997; Walter et al. 2000) that some authors
water in the terrestrial planets. Unfortunately, we do not know      believed would effectively degas the planets of any existing
the original abundance and isotopic composition of water on          water.
Venus because of the solar UV water dissociation loss                     We now know that there are elemental and isotopic
mechanism. It is possible that we do not know the intrinsic          reasons why at best 50% and most probably a very small
isotopic composition of water on Mars because Mars lacks             percentage of water accreted to Earth from cometary impacts
plate tectonics and Martian meteorites may simply be                 (Drake and Righter 2002). Figure 3 compares the isotopic
sampling water delivered subsequent to planetary formation,          composition of hydrogen in Earth, Mars, three Oort Cloud
although Leshin-Watson et al. (1994) argue to the contrary—          comets, and various early solar system estimates. It is
that SNC meteorites provide evidence for an intrinsic Martian        immediately clear that 100% of Earth’s water did not come
mantle D/H ratio up to 50% higher than Earth’s mantle. So we         from the Oort Cloud comets. D/H ratios in Martian meteorites
must look to the best-studied planet, Earth, for clues. But first,   do agree with cometary values, but that may simply reflect the
let us look at the pros and cons of various proposed sources of      impact of comets onto the Martian surface in a non-plate
water.                                                               tectonics environment that precludes recycling.
Origin of water in the terrestrial planets
Origin of water in the terrestrial planets                                            3

Fig. 2. Distribution of water ice on Mars, inferred from the Mars Odyssey Gamma Ray Spectrometer instrument (Boynton et al. 2002). Ice
sheets at least 1 m thick are covered by a thin layer of dust polewards of approximately 60° of latitude in both hemispheres.

     So what limits the cometary contribution to Earth’s            as has been confirmed in preliminary laboratory experiments
water? Consider, for example, that perhaps Earth accreted           on pure water ice (Weirich et al. 2004). Lower bulk D/H ratios
some hydrous phases or adsorbed water, and some amount of           would increase the allowable amount of cometary water.
additional water came from comets. Indigenous Earth water           Intriguing experiments on mixtures of water ice and TiO2
could have had D/H ratios representative of the inner solar         grains by Moores et al. (2005) suggest that D/H ratios could
system, i.e., low values because of relatively high nebular         be lowered in sublimates. Third, the D/H ratios of organics
temperatures, perhaps like protosolar hydrogen (2–3 × 10−5,         and hydrated silicates in comets are unknown, although that
Lecluse et al. 1994), in which case a cometary contribution of      situation may be rectified by analysis of samples returned by
up to 50% is possible. Alternatively, indigenous Earth water        the Stardust mission. Note, however, that D/H ratios up to 50×
could have had D/H ratios representative of a protosolar water      Vienna Standard Mean Ocean Water (VSMOW) have been
component identified in meteorites (~9 × 10−5, Deloule et al.       measured in some chondritic porous interplanetary dust
1995), in which case there could be as little as a 10–15%           particles (CP-IDPs) that may have cometary origins
cometary contribution (Owen and Bar-nun 2000).                      (Messenger 2000), and higher aggregate D/H ratios of comets
     There are caveats to using cometary D/H ratios to limit        would decrease the allowable cometary contribution to
the delivery of cometary water to Earth. First, we do not know      Earth’s water.
if the Oort Cloud comets Halley, Hale-Bopp, and Hyakutake                Delivery of water from comets can also be evaluated in
are representative of all comets. Certainly, they are unlikely to   light of other cometary geochemical data. For instance, for an
be representative of the Kuiper Belt objects, the source of         assumed Ar/H2O ratio of 1.2 × 10−7 in the bulk Earth, comets
Jupiter-family comets, as the Oort Cloud comets formed in           like Hale-Bopp with an approximately solar ratio of Ar/H2O
the region of the giant planets and were ejected, while Kuiper      (Stern et al. 2000) would bring in 2 × 104 more Ar than is
Belt objects have always resided beyond the orbit of Neptune.       presently found in the Earth’s atmosphere (Swindle and Kring
Second, D/H measurements are not made of the solid nucleus,         2001), if 50% of Earth’s water—the maximum amount
but of gases emitted during sublimation. Differential diffusion     permitted by D/H ratios—were derived from comets. It is
and sublimation of HDO and H2O may make such                        unclear if this measurement of comet Hale-Bopp is applicable
measurements unrepresentative of the bulk comet, as the D/H         to all comets. Ar/O ratios of
Origin of water in the terrestrial planets
4                                                                 M. J. Drake

                                                                          Asteroids

                                                                               Asteroids are a plausible source of water based on
                                                                          dynamical arguments. Morbidelli et al. (2000) have shown
                                                                          that up to 15% of the mass of the Earth could be accreted late
                                                                          in Earth’s growth by collision of one or a few asteroids
                                                                          originating in the Main Belt. However, there are strong
                                                                          geochemical arguments against a significant contribution of
                                                                          water from asteroids, unless one postulates that Earth was hit
                                                                          by a hydrous asteroid unlike any falling to Earth today, as
                                                                          sampled by meteorites. One cannot prove this hypothesis
                                                                          wrong, as it could involve a single, unique event.
                                                                               However, if asteroidal material falling to Earth 4.5 Ga
                                                                          ago was the same as that which falls today, one can effectively
Fig. 3. The D/H ratios in H2O in three comets, meteorites, Earth          rule out asteroids as a source of water. The reason involves
(Vienna Standard Mean Ocean Water [SMOW]), protosolar H2, and
                                                                          the Os isotopic composition of the so-called “late veneer,” the
Mars. CC = carbonaceous chondrites, LL3-IW = interstellar water in
Semarkona, LL3-PS = protostellar water in Semarkona. After Drake          material that may have contributed the highly siderophile
and Righter (2002).                                                       elements (HSEs) that are present to within 4% of chondritic
                                                                          proportions at about 0.003 of chondritic absolute abundances
                                                                          (see Fig. 1 of Drake and Righter 2002). Note that the term
                                                                          “late veneer” is misleading but is firmly entrenched in the
                                                                          literature. A veneer is a surface coating, yet the “late veneer”
                                                                          of HSEs must be well mixed into at least the upper mantle of
                                                                          Earth and perhaps the entire silicate mantle.
                                                                               Of current meteorite falls, only carbonaceous chondrites
                                                                          have significant amounts of water (Drake and Righter 2002).
                                                                          Enstatite and ordinary chondrites are anhydrous. The late
                                                                          addition of water must be accompanied by other chemical
                                                                          elements such as Re and Os. Rhenium and Os are two HSEs
                                                                          that are linked by beta-decay, 187Re = 187Os + β, thus Os
                                                                          isotopes can be used as a constraint on Earth’s bulk
                                                                          composition. Earth’s primitive upper mantle (PUM) has a
                                                                          significantly higher 187Os/188Os ratio than carbonaceous
                                                                          chondrites, effectively ruling them out as the source of the
                                                                          “late veneer” (Fig. 4). The PUM 187Os/188Os ratio overlaps
Fig. 4. 187Os/188Os ratios in carbonaceous, ordinary, and enstatite       anhydrous ordinary chondrites and is distinctly higher than
chondrites, and in the Earth’s primitive upper mantle (PUM), are          anhydrous enstatite chondrites.
distinct and are diagnostic of the nature of the Earth’s “late veneer.”
                                                                               The identification of anhydrous meteorites with the “late
Mars is not plotted because the uncertainty in its initial 187Os/188Os
ratio is larger than range of the x-axis. After Drake and Righter         veneer” (Walker et al. 2002) effectively rules out asteroidal
(2002).                                                                   material being the source of Earth’s water. There is one
                                                                          caveat, however. Thermal processing of asteroids was
LINEAR 2000 WM1 (Weaver et al. 2002). However, the true                   occurring 4.5 Ga ago. One cannot exclude the possibility that
Ar/O ratios in comets would have to be at least three orders of           ordinary chondrites once contained water and those falling to
magnitude below solar in order to be consistent with the Ar               Earth today have lost it by metamorphism, even though water
abundance of the Earth’s atmosphere.                                      was still present 4.5 Ga ago. However, the preservation of
     Another approach to estimating the contribution of                   aqueous alteration products in some carbonaceous chondrites
cometary materials to Earth’s water budget can be made by                 (McSween 1979) suggests that loss of water from initially
considering the implications for the abundances of noble                  hydrous asteroids is unlikely to proceed to the anhydrous
metals and noble gases. Dauphas and Marty (2002) show that                limit.
the total mass of cometary and asteroidal material accreting to
Earth after core formation is 0.7–2.7 × 1025 g and that comets            Phyllosilicates
contribute
Origin of water in the terrestrial planets
Origin of water in the terrestrial planets                                            5

phyllosilicates from the asteroid belt, where they clearly exist.    to the terrestrial planets exceeded the mass of water accreted
This hypothesis accepts that the region of the accretion disk        by a factor of about 12 if all four terrestrial planets initially
where the terrestrial planets reside was too hot for hydrous         had 50 Earth oceans of water. This excess represents a
phases to exist. This hypothesis appears subject to the same         minimum in that the water vapor content of the accretion disk
objection as other asteroidal sources of water. It seems             inside 3 AU could exceed 2 Earth masses, and it is unlikely
unlikely that phyllosilicates could be decoupled from other          that all four terrestrial planets received 50 Earth oceans. There
minerals and transported into the inner solar system. Thus, the      is also the possibility of enhancing the available water vapor
Os isotopic composition of PUM is a constraint in this case as       reservoir by inward migration of ices (Krot et al. 2004) and
well.                                                                phyllosilicates (Ciesla et al. 2004).
     However, a variant on the hypothesis may be worth                    The question is the fate of this water vapor. Let us accept
exploring. Inward migration of phyllosilicates is postulated to      for the purposes of discussion that it was too hot inside 3 AU,
occur in an environment where nebular gas was still present          even as the accretion disk cooled, for hydrous minerals to
and temperatures were too high for hydrated minerals such as         form. Could water vapor be adsorbed onto grains before the
phyllosilicates to be thermodynamically stable. Thus, if a way       gas in the inner solar system was dissipated? This question
could be found to migrate phyllosilicates either by themselves       requires knowledge of grain volume distribution, fractal
or embedded with other minerals such that their contribution         dimensions of grains, temperature (which affects the velocity
to the Os isotopic composition of PUM was insignificant,             with which molecules impact grains and hence the molecules’
dehydration of phyllosilicates could enhance the water vapor         “stickiness”), competition between water molecules and
content of the region of the terrestrial planets. Perhaps some       others (H2 is the dominant gas molecule, followed by He), and
of the phyllosilicates were accreted to the terrestrial planets as   retentiveness after adsorption (can a molecule be knocked off
Ciesla et al. (2004) propose without perturbing the Os               the grain by impact of another molecule?). This question also
isotopic composition of PUM, but the majority of water               requires knowledge of whether the molecules are adsorbed by
carried in by them was adsorbed onto grains already present          low energy physisorption, or higher energy chemisorption.
in the region of the terrestrial planets and incorporated into       Physisorption and chemisorption span the range of energies
the terrestrial planets as they grew as discussed below (Drake       from van der Waals forces around 5 kJ/mol through strong
2004; Stimpfl et al. 2004).                                          hydrogen bonding around 100 kJ/mol, just slightly weaker
                                                                     than covalent bonds. Various types of O-H bonds span this
 COULD THE TERRESTRIAL PLANETS ACCRETE                               range of energies (Jeffrey 1997).
                “WET”?
                                                                     Physisorption
     The solar system formed as a result of the collapse of a
large, cold, slowly rotating cloud of gas and dust into a disk            Stimpfl et al. (2004) have examined the role of
that defined the plane of the solar system. The terrestrial          physisorption by modeling the adsorption of water at 1000 K,
planets grew in this accretion disk. Hydrogen, helium, and           700 K, and 500 K using a Monte Carlo simulation with a grid
oxygen dominated the gas, in which the dust was bathed.              of 10,000 adsorption sites. The composition of the nebular
Some of that hydrogen and oxygen combined to make water              gas was 99.98 wt% H2 and He, with an H2O/H2 ratio of 5.4 ×
vapor. If thermodynamic equilibrium was attained, Lecluse            10−4 (Lodders 2003). All of the gases (H2, He, and H2O)
and Robert (1994) estimated that there were about three Earth        interacted with the surface, but only H2O could be adsorbed
masses of water vapor in the accretion disk inside of 3 AU.          (Stimpfl, personal communication). An iterative process
Subsequent spectroscopic work has led to a downward                  allowed the surface to reach steady state saturation at each
revision of the carbon abundance by 30% (Palme and Jones             temperature. Water molecules not only interact with the
2003) and the oxygen abundance by 40% (Lodders 2003).                substrate by means of weak bonds (~5 kJ/mol) but also
Thus, it now appears that there were about two Earth masses          establish hydrogen bonds with other water molecules present
of water vapor in the accretion disk inside of 3 AU. There are       in a monolayer (de Leeuw et al. 2000). Stimpfl et al. (2004)
arguments associated with the photodissociation of CO, the           took this cooperative behavior into account by increasing the
dominant C-bearing gas in the accretion disk, that suggest the       bond energy proportionally to the number of nearest
water abundance could have been modestly higher than the             neighbors (max. allowed = 4). The energy of the incoming
value calculated from simple equilibrium thermodynamics              molecules was computed using the Maxwell-Boltzmann
(Clayton 2004). The mass of the Earth is 5 × 1027 g. The mass        probability distribution. Simple “sticking” rules were applied.
of one Earth ocean is 1.4 × 1024 g. The extreme maximum              Incoming water molecules stuck to the surface if their kinetic
estimate for the amount of water in the Earth is about 50 Earth      energy was lower than 5 kJ/mol. If any molecule collided
oceans (Abe et al. 2000), with most estimates being 10 Earth         with a molecule already occupying a site, the resident
oceans or less. For example, an estimate based on the water          molecule was dislodged only if the incoming molecule had an
storage potential of minerals in the silicate Earth is about 5–6     energy exceeding two times the total bond energy of the
Earth oceans (Ohtani 2005). Thus, the mass of water available        bonded molecule. Stimpfl et al. (2004) allowed only for the
6                                                              M. J. Drake

Fig. 5. Results of ab initio calculations at 0 K, in which the Gibbs free energy of Si-O-H clusters is minimized and then an H2O molecule is
introduced (Gibbs, personal communication 2004). Silicon is pale blue, oxygen is red, original hydrogens are black, and hydrogens attached
to the introduced water molecule are dark blue. a) The Gibbs Free Energy (G) of a Si-O-H assemblage is minimized. b) A water molecule is
introduced in the vicinity of the minimum G Si-O-H assemblage. c) A thermodynamically unstable early bonding of the water molecule with
the Si-O-H assemblage. d) The final minimized G state.

adsorption of one monolayer, neglected porosity and surface            hence making water more retentive. I address chemisorption
roughness, considered water to be an infinite reservoir, and           next.
assumed that all the particles interacting with the surface
were water molecules.                                                  Chemisorption
     Stimpfl et al. (2004) “exploded” the Earth into 0.1 m
spheres of volume equal to Earth, initially ignoring the fractal            Calculations of chemisorption of water onto accretion
nature of these particles. The results were disappointing. Only        disk grains at realistic nebular temperatures have not been
0.25% of an ocean of water could be adsorbed at 1000 K, 1%             performed. However, Gibbs has conducted ab initio
of one Earth ocean could be adsorbed at 700 K, and only 3%             calculations on the docking sites of H onto SiO2 polymorphs
at 500 K.                                                              (e.g., Gibbs et al. 2004). Gibbs (personal communication) has
     However, grains in the accretion disk were not spherical,         also conducted ab initio calculations at 0 K, in which the
but fractal. They were also not of uniform volume. Thus, the           Gibbs free energy of Si-O-H clusters is minimized and then a
available surface area for adsorption is vastly greater than that      H2O molecule is introduced. Dynamical visualization shows
of a sphere of corresponding volume. If the surface area of the        that, after passing through several metastable bonding states
fractal grain was 100 times that of a sphere of corresponding          and when the minimum Gibbs free energy is achieved, strong
volume, then one quarter of an ocean of water could be                 bonds (>>5 kJ/mol) can be formed between the water
adsorbed at 1000 K, one Earth ocean could be adsorbed at               molecule and the Si-O-H cluster (Fig. 5). If this strong
700 K, and three Earth oceans could be adsorbed at 500 K.              bonding behavior persists to realistic nebular temperatures,
     Of course, there are issues of retention of water as the          retention of adsorbed water during the later violent stages of
grains collide and grow to make planets. Stimpfl et al. (2004)         accretion is enhanced. This is an active area of current study
note that some of this water will be lost due to high collisional      by Gibbs.
velocities, especially after the grains grow to a radius >10 cm.
However, it is clear that volatiles are not completely                              MAJOR UNSOLVED QUESTIONS
outgassed even in planetary-scale collisions. For example,
primordial 3He, far more volatile than water, is still                      We have just started examining the idea that the
outgassing from Earth’s mantle 4.5 Ga after an almost grown            terrestrial planets could have obtained much or all of their
Earth collided with a Mars-sized body, which gave origin to            water by adsorption onto grains directly from gas in the
the Moon.                                                              accretion disk, and many major issues remain to be resolved.
     Other processes may counter the tendency for loss of              Below I list some of the most important.
water. For example, the temperature at any given heliocentric            1. The D/H ratio of the nebular gas is inferred from
distance in the accretion disk decreases as a function of time.             spectroscopic measurements of CH4 in the atmospheres
Lower temperatures may favor water being chemisorbed, a                     of Jupiter and Saturn to be 2.1 ± 0.4 × 10−5 (Lellouch et
process involving higher bond energies than ~5 kJ/mol and                   al. 2001), much lower than VSMOW. Spectroscopic or
Origin of water in the terrestrial planets                                             7

       direct measurements of solar D/H ratio cannot be made          700 K, the efficiency of adsorption of water increases as
       because practically all of the Sun’s deuterium has been        temperature decreases. Thus, the process should have been
       burned to make 3He. If the nebular D/H ratio really is as      more efficient further from the Sun than closer to the Sun.
       low as implied by the Jovian and Saturnian atmospheres,        Thus, it is likely that Mars, Earth, and Venus all accreted some
       a mechanism to raise the D/H ratio of nebular gas from         water by adsorption, with Mars accreting the most, both
       solar to VSMOW is needed.                                      because of its greater distance from the Sun and the lower
  2.   It is likely that the D/H and Ar/O ratios measured in          energy of collisions during accretion because of its smaller
       cometary comas and tails are not truly representative of       final mass. (A caveat—Mars’ small size might be the result of
       cometary interiors. As noted earlier, D/H ratios in            higher energy collisions because of the gravitational pumping
       laboratory experiments can increase or decrease with           of relative velocities of planetesimals by Jupiter.) Mars
       time due to differential diffusion and sublimation,            accreting more water than Earth is consistent with Mars’
       depending on the physical nature of the starting material      enrichment in volatile lithophile elements (Dreibus and
       (Weirich et al. 2004; Moores et al. 2005). Depending on        Wänke 1985). Even Mercury may have accreted some water
       the siting of Ar in comets, Ar/O ratios may also be            by this mechanism, even though there is no evidence for water
       similarly affected. The Deep Impact mission may resolve        at its surface today. Therefore, the current differences in the
       this issue by exposing fresh cometary interior material        apparent water abundances among the terrestrial planets are
       for spectral analysis with ground-based and space-based        probably the result of both different initial inventories and
       high spectral resolution spectrometers.                        subsequent geologic and atmospheric processing.
  3.   The key argument against an asteroidal source of Earth’s             It is widely believed that the Moon formed in a giant impact
       water is that the Os isotopic composition of Earth’s           (Canup and Asphaug 2001). Drake and Righter (2002) noted
       primitive upper mantle matches that of anhydrous               that the identical oxygen isotopic composition of Earth and
       ordinary chondrites, not hydrous carbonaceous                  Moon (Wiechert et al. 2001) imply that the impactor was made
       chondrites. But are the parent bodies of the ordinary          of the same material that formed Earth. If the Earth accreted
       chondrites anhydrous? Could ordinary chondritic                “wet,” then so did the impactor. Today the Moon appears to be
       meteorites be derived from the metamorphosed outer             devoid of water, at least to depths of approximately 500 km, as
       parts of hydrous asteroids, in which case impact of a bulk     sampled by mare basalt glasses (Delano 1980). Thus, the fact
       asteroid could deliver water? It is probable that spacecraft   that the Moon is bone-dry today must be a consequence of the
       spectral examination of very deep impact basins in S-type      physics and chemistry occurring during the extreme
       asteroids will be needed to address this question.             environments experienced by the ejecta following the giant
  4.   A related question is why there are any anhydrous              impact. On a much smaller scale, tektites are anhydrous even
       primordial bodies in the solar system if adsorption of         though most formed during impacts into hydrous granitic
       water from gas in the accretion disk was an efficient          terrain (Fudali et al. 1987; Brett and Sat 1984).
       process, as preliminary calculations suggest it might                If the terrestrial planets were initially hydrous, there are
       have been.                                                     important implications for the late stages of accretion and
  5.   The timing of loss of gas from the accretion disk in the       primordial planetary differentiation. Water depresses the
       region of the terrestrial planets is unknown. For              liquidus and solidus of magmas (see Fig. 8 of Righter and
       adsorption to be efficient, nebular gas must have              Drake [1999] for a summary diagram for peridotite). For a
       persisted long after the initial aggregation of grains into    water-saturated peridotite, that depression can be as much as
       planetesimals. The timing of loss of gas from the              700 °C to 1000 °C, depending on pressure. While it is
       accretion disk will be intimately connected to the             unlikely that the terrestrial planets accreted enough water to
       currently unknown mechanism of loss.                           have water-saturated magmas, “wet” accretion implies much
                                                                      deeper planetary magma oceans than would be the case for
                        IMPLICATIONS                                  ahydrous planets for the same impact energy.
                                                                            There is another interesting consequence for the evolution
     The idea that planetary water could have been obtained           of planetary redox states if the terrestrial planets accreted
by adsorption of nebular water onto grains in the accretion           “wet.” Okuchi (1997) showed that when Fe-metal and water
disk and retained during planetary accretion remains                  were compressed to 75 kbars and heated to 1200–1500 °C,
unproven because quantitative studies have just begun. In that        iron hydride formed. Thus, planetary cores should contain H,
sense, my Leonard Medal Address is risky but provocative              and OH should be left behind in the molten silicate. As more
because early results (Stimpfl et al. 2004) indicate that the         metal was delivered to the planet while it accreted, more H
idea has promise. If some or all of Earth’s water was obtained        would be extracted into the core and more OH would be
by adsorption, then there are number of interesting                   liberated in the silicate. Thus, planetary mantles should
implications.                                                         become progressively oxidized with time, perhaps explaining
     Stimpfl et al. (2004) showed that, while it is plausible to      the high redox states relative to the iron-wüstite buffer.
adsorb 1–3 Earth oceans of water at temperatures of 500 K to                It remains possible that the “late veneer” may really be of
8                                                           M. J. Drake

asteroidal origin. If so, it has the composition of ordinary        whose published contributions have contributed to the
chondrites and cannot be the source of water delivered after        advancement of our knowledge of meteoritics specifically
the bulk of the terrestrial planets accreted. Late addition of up   and planetary science in general. These individuals, in
to 5 wt% asteroidal material is consistent with dynamical           chronological order of joining my group, are Richard Bild,
simulations (Morbidelli et al. 2000), though the “late veneer”      Duck Mittlefehldt, John Jones, Jack Berkley, Horton
appears to represent less than 1 wt% of the mass of the Earth.      Newsom, Charles Hostetler, Allan Treiman, Dan Malvin, Lisa
However, Dauphas et al. (2004) use Ru and Mo isotopic               McFarlane, Valerie Hillgren, Leigh Broadhurst, Cyrena
systematics to argue against this possibility and argue for         Goodrich, Chris Capobianco, Don Musselwhite, Kevin
homogeneous accretion. They note that most Mo in the                Righter, Nancy Chabot, Windy Jaeger, Werner Ertel, Ken
Earth’s mantle is delivered while core formation is occurring       Domanik, Steve Singletary, and Marilena Stimpfl. In a very
because Mo is moderately siderophile, while essentially all         real sense, the Leonard Medal is their achievement too. Jerry
Ru must be delivered after core formation because it is highly      Gibbs is thanked for allowing the use of visualizations of his
siderophile. Thus, their isotopic systematics must record the       unpublished ab initio calculations. Comments by Andy Davis
first 99% of accretion (Mo) and the last 1% of accretion (Ru).      and Conel Alexander during questioning following the
The fact that the silicate Earth falls on the cosmic Mo-Ru          Leonard Medal Address are appreciated and views
correlation argues that the bulk of the Earth and the “late         incorporated. A review by Andy Davis corrected a glaring
veneer” accreted from the same Mo-Ru isotopic reservoir.            error and is much appreciated. This work was supported by
      Finally, the evidence that the Earth and other terrestrial    NASA grant NAG5-12795.
planets accreted “wet” has implications for the first
appearance of liquid water oceans. There is intriguing              Editorial Handling—Dr. A. J. Timothy Jull
evidence for widespread water on Earth and Mars within
about 100 Myr of the nucleosynthesis of 129I. Earth and Mars                                 REFERENCES
preserve reservoirs with distinct 129Xe/132Xe ratios. The
isotope 129Xe is produced through the decay of now extinct          Abe Y., Ohtani E., Okuchi T., Righter K., and Drake M. J. 2000.
129I (t1/2 = 16 Myr). After 5–7 half-lives, radiogenic daughter         Water in the early Earth. In Origin of the Earth and Moon, edited
129Xe from the decay of 129I cannot be detected. It is difficult,       by Canup R. M. and Righter K. Tucson: The University of
                                                                        Arizona Press. pp. 413–434.
although not impossible, to fractionate I from Xe by purely
                                                                    Baker V. R., Strom R. G., Gulick V. C., Kargel J. S., Komatsu G., and
magmatic processes (Musselwhite and Drake 2000).                        Kale V. S. 1991. Ancient oceans, ice sheets and the hydrological
However, water is far more effective at fractionating I from            cycle on Mars. Nature 352:589–593.
Xe (Musselwhite et al. 1991) because I dissolves in liquid          Boss A. P. 1998. Temperatures in protoplanetary disks. Annual
water (it is a halogen) while Xe bubbles through. If accretion          Reviews of Earth and Planetary Science 26:26–53.
                                                                    Boynton W. V., Feldman W. C., Squyres S. W., Prettyman T. H.,
ceased while 129I still existed and any magma ocean
                                                                        Bruecher J., Evans L. G., Reedy R. C., Starr R., Arnold J. R.,
solidified, liquid water could become stable at planetary               Drake D. M., Englert P. A. J., Metzger A. E., Mitrofanov I.,
surfaces. Iodine (including still extant 129I) dissolved in a           Trombka J. I., d’Uston C., Wänke H., Gasnault O., Hamara D. K.,
liquid water ocean would react hydrothermally with any                  Janes D. M., Marcialis R. L., Maurice S., Mikheeva I., Taylor G.
crust. Recycling of 129I into the mantles of Earth and Mars             J., Tokar R., and Shinohara C. 2002. Distribution of hydrogen in
                                                                        the near surface of Mars: Evidence for subsurface ice deposits.
would lead to their mantles having excess 129Xe. The distinct
129Xe/132Xe ratios in different reservoirs in Earth and Mars
                                                                        Science 297:81–85.
                                                                    Brett R. and Sato M. 1984. Intrinsic oxygen fugacity measurements
can be developed by outgassing of the mantles, coupled with             on seven chondrites, a pallasite, and a tektite and the redox state
atmospheric stripping in the case of Mars (Musselwhite 1995;            of meteorite parent bodies. Geochimica et Cosmochimica Acta
Musselwhite et al. 1991). Thus, it is possible that Venus,              48:111–120.
                                                                    Canup R. M. and Asphaug E. 2001. The Moon-forming impact.
Earth, and Mars all had liquid water oceans shortly after their
                                                                        Nature 412:708–712.
magma ocean epochs.                                                 Carr M. H. 1996. Water on Mars. New York: Oxford University
                                                                        Press. 229 p.
Acknowledgments–The Leonard Medal is given to a senior              Ciesla F. J., Lauretta D. S., and Hood L. L. 2004. Radial migration of
member of the community for lifetime achievement. Except                phyllosilicates in the solar nebula (abstract #1219). 35th Lunar
in rare cases, the contributions to scholarship were not                and Planetary Science Conference. CD-ROM.
                                                                    Clayton R. N. 2004. Oxygen isotopic compositions of the terrestrial
singular affairs, but the product of collaborations with many           planets. Workshop on Oxygen in the Terrestrial Planets. LPI
talented individuals who joined the research group. For                 Contribution #1203. Houston: Lunar and Planetary Institute. p. 16.
guidance in my formative years as a scientist, I owe enormous       Dauphas N. and Marty B. 2002. Inference on the nature and mass of
debts of gratitude to William S. Fyfe, my undergraduate                 Earth’s late veneer from noble metals and gases. Journal of
mentor, to my Ph.D. advisor Daniel S. Weill, and to my                  Geophysical Research 107:E12-1–E12-7.
                                                                    Dauphas N., Davis A. M., Marty B., and Reisberg L. 2004. The
postdoctoral advisor, John A. Wood. My postdoctoral year                cosmic molybdenum-ruthenium isotope correlation. Earth and
sharing an office with Jeff Taylor was memorable. I also wish           Planetary Science Letters 226:465–475.
to acknowledge members of my research group over the years          Delano J. W. 1980. Chemistry and liquidus phase relations of Apollo
Origin of water in the terrestrial planets                                                    9

    15 red glass: Implications for the deep lunar interior.                Morbidelli A., Chambers J., Lunine J. I., Petit J. M., Robert F.,
    Proceedings, 11th Lunar and Planetary Science Conference. pp.              Valsecchi G. B., and Cyr K. 2000. Source regions and time scales
    251–288.                                                                   for delivery of water to the Earth. Meteoritics & Planetary
de Leeuw N. H., Parker S. C., Catlow C. R. A., and Proce G. D. 2000.           Science 35:1309–1320.
    Modeling the effect of water on the surface structure and stability    Musselwhite D. S. 1995. Experimental geochemistry of iodine,
    of forsterite. Physics and Chemistry of Minerals 27:332–341.               argon, and xenon: Implications for the early outgassing histories
Deloule E. and Robert F. 1995. Interstellar water in meteorites?               of the Earth and Mars. Ph.D. thesis, The University of Arizona,
    Geochimica et Cosmochimica Acta 59:4695–4706.                              Tucson, Arizona, USA.
Donahue T. M. and Pollack J. B. 1983. Origin and evolution of the          Musselwhite D. S., Drake M. J., and Swindle T. D. 1991. Early
    atmosphere of Venus. In Venus, edited by Hunten D. M., Colin L.,           outgassing of Mars: Inferences from the geochemistry of iodine
    and Donahue T. M. Tucson: The University of Arizona Press. pp.             and xenon. Nature 352:697–699.
    1003–1036.                                                             Musselwhite D. S. and Drake M. J. 2000. Early outgassing of Mars:
Drake M. J. 2004. Origin of water in the terrestrial planets (abstract).       Implications from experimentally determined solubility of iodine
    Meteoritics & Planetary Science 39:A31.                                    in silicate magmas. Icarus 148:160–175.
Drake M. J. and Righter K. 2002. Determining the composition of the        Ohtani E. 2005. Water in the mantle. Elements 1:25–30.
    Earth. Nature 416:39–44.                                               Okuchi T. 1997. Hydrogen partitioning into molten iron at high
Dreibus G. and Wänke H. 1985. Mars, a volatile-rich planet.                    pressure: Implications for Earth’s core. Science 278:1781–1784.
    Meteoritics 20:367–381.                                                Owen T. and Bar-nun A. 2000. Volatile contributions from icy
Fudali R. F., Dyar M. D., Griscom D. L., and Schreiber H. D. 1987.             planetesimals. In Origin of the Earth and Moon, edited by Canup
    The oxidation state of iron in tektite glass. Geochimica et                R. M and Righter K. Tucson: The University of Arizona Press.
    Cosmochimica Acta 51:2749–2756.                                            pp. XX–XX.
Gibbs G. V., Cox D. F., and Ross N. L. 2004. A modeling of the             Palme H. and Jones A. 2003. Solar system abundances of the
    structure and favorable H-docking sites and defects for the high-          elements. In Treatise on geochemistry, vol. 1., edited by Holland
    pressure silica polymorph stishovite. Physics and Chemistry of             H. D. and Turekian K. K. Amsterdam: Elsevier. pp. 41–61.
    Minerals 31:232–239.                                                   Righter K. and Drake M. J. 1996. Core formation in Earth’s Moon,
Gradie J. C., Chapman C. R., and Tedesco E. F. 1989. Distribution of           Mars, and Vesta. Icarus 124:513–529.
    taxonomic classes and the compositional structure of the asteroid      Righter K. and Drake M. J. 1997. Metal/silicate equilibrium in a
    belt. In Asteroids II, edited by Binzel R. P, Gehrels T., and              homogeneously accreting Earth: New results for Re. Earth and
    Matthews M. S. Tucson: The University of Arizona Press. pp.                Planetary Science Letters 146:541–554.
    316–335.                                                               Stern S. A., Slater D. C., Festou M. C., Parker J. W., Gladstone G. R.,
Harmon J. 1997. Mercury radar studies and lunar comparisons.                   A’Hearn M. F., and Wilkinson E. 2000. The discovery of argon
    Advances in Space Research 19:1487–1496.                                   in Comet C/1995 01 (Hale-Bopp). The Astrophysical Journal
Jeffrey G. A. 1997. An introduction to the hydrogen bond. New York:            544:L169–L172.
    Oxford University Press. 303 p.                                        Stimpfl M., Lauretta D. S., and Drake M. J. 2004. Adsorption as a
Krot A. N., Hutcheon I. D., Yurimoto H., Cuzzi J. N., McKeegan K.              mechanism to deliver water to the Earth (abstract). Meteoritics &
    D., Scott E. R. D., Libourel G., Chaudisson M., Aléon J., and              Planetary Science 39:A99.
    Petaev M. I. Forthcoming. Evolution of oxygen isotopic                 Swindle T. D. and Kring D. A. 2001. Implications of noble gas
    composition in the inner solar nebula. The Astrophysical Journal.          budgets for the origin of water in Earth and Mars (abstract
Lecluse C. and Robert F. 1994. Hydrogen isotope exchange reaction              #3785). Eleventh Annual V. M. Goldschmidt Conference. CD-
    rates: Origin of water in the inner solar system. Geochimica et            ROM.
    Cosmochimica Acta 58:2927–2939.                                        Taylor S. R. 2001. Solar system evolution: A new perspective, 2nd ed.
Lellouch E., Bezard B., Fouchet T., Feuchtgruber H., Encrenaz T.,              Cambridge: Cambridge University Press. 460 p.
    and de Graauw T. 2001. The deuterium abundance of Jupiter and          Walker R. J., Horan M. F., Morgan J. W., Becker H., Grossman J. N.,
    Saturn from ISO-SWS observations. Astronomy & Astrophysics                 and Rubin A. E. 2002. Comparative 187Re-187Os systematics of
    370:610–622.                                                               chondrites: Implications regarding early solar system processes.
Leshin-Watson L., Hutcheon I. D., Epstein S., and Stolper E. M.                Geochimica et Cosmochimica Acta 66:4187–4201.
    1994. Water on Mars: Clues from deuterium/hydrogen and water           Walter M. J., Newsom H., Ertel W., and Holzheid A. 2000.
    contents of hydrous phases in SNC meteorites. Science 265:86–              Siderophile elements in the Earth and Moon: Metal-silicate
    90.                                                                        partitioning and implications for core formation. In Origin of the
Li J. and Agee C. 1996. Geochemistry of mantle-core formation at               Earth and Moon, edited by Canup R. M. and Righter K. Tucson:
    high pressure. Nature 381:686–689.                                         The University of Arizona Press. pp. XX–XX.
Lodders K. 2003. Solar system abundances and condensations                 Wiechert U., Halliday A. N., Lee D-C., Snyder G., Taylor L. A., and
    temperatures of the elements. The Astrophysical Journal 591:               Rumble D. 2001. Oxygen isotopes and the Moon-forming giant
    1220–1247.                                                                 impact. Science 294:345–348.
McSween H. Y., Jr. 1979. Are carbonaceous chondrites primitive or          Weaver H. A., Feldman P. D., Combi M. R., Krasnopolsky V., Lisse
    processed? A review. Reviews of Geophysics and Space Physics               C. M. and Shermansky D. E. 2002. A search of Ar and O VI in
    17:1059–1078.                                                              three comets using the far ultraviolet spectroscopic explorer. The
Messenger S. 2000. Identification of molecular cloud material in               Astrophysical Journal 576:L95–L98.
    interplanetary dust particles. Nature 404:968–971.                     Weirich J. R., Brown R. H., and Lauretta D. S. 2004. Cometary D/H
Moores J. E., Brown R. P., Lauretta D. S., and Smith P. H. 2005.               fractionation during sublimation (poster). 36th Annual Meeting
    Preliminary results of sublimations fractionation in dusty                 of the Division for Planetary Sciences of the American
    disaggregated samples (abstract #1973). 36th Lunar and                     Astronomical Society. Bulletin of the American Astronomical
    Planetary Science Conference. CD-ROM.                                      Society 36:1143.
You can also read